采用Fenton氧化法进行人工智能芬顿氧化处理间甲酚废水实验,考察了Fe²⁺质量浓度、H₂O₂体积分数、初始pH、反应时间和间甲酚初始质量分数对降解间甲酚反应的影响,利用响应面法(RSM)和人工神经网络(ANN)分别确定降解间甲酚的最佳方案,同时对TOC去除率的关系进行拟合优化对比。结果表明,利用ANN模型并采用枚举法获取的最佳优化条件:Fe²⁺质量浓度为0.66 g/L、H₂O₂体积分数为6.00 mL/L、初始pH为3、反应时间为23.37 min、间甲酚初始质量分数为50 μg/g,此时,TOC去除率为48.14%,优于响应面法的32.16%。

m-Cresol-containing wastewater is treated by using artificial intelligence assisted Fenton oxidation method.The impacts of Fe²⁺ mass concentration,H₂O₂ volume fraction,initial pH,reaction time,and initial m-cresol mass fraction on m-cresol degradation are explored.The best scheme for degradation of M-cresol is determined by using response surface methodology (RSM) and artificial neural networks (ANN),respectively while the relationship with TOC removal rate is fitted to optimize and compare.The optimal conditions obtained by the ANN model combined with an enumerative method are as follows:Fe²⁺ mass concentration is 0.66 g·L^-1,H₂O₂ volume fraction is 6.00 mL·L^-1,initial pH is 3,reaction time is 23.37 min,and the initial mass fraction of m-cresol is 50 μg·g^-1.Under these conditions,the removal rate of TOC is 48.14%,that is 32.16% under the conditions obtained via the response surface method.

张婧(1998-),女,硕士生,研究方向为人工智能与水处理,2021520037@bipt.edu.cn;马磊(1986-),男,工学博士,副教授,研究方向为催化氧化的方法处理难降解工业废水,通讯联系人,malei@bipt.edu.cn。